Phylogenetic comparative methods were used to test hypotheses about cultural evolution in ethnolinguistic groups from the Austronesian language family of the Pacific. The case for quantitative statistical approaches to the empirical evolution of linguistic and cultural features was presented. Phylogenetic trees of 67 Austronesian languages were constructed using maximum parsimony and Bayesian Markov Chain Monte Carlo likelihood algorithms on a database of lexical items. The predominant transmission mode of 76 cultural traits was examined at the macroevolutionary level with (i) partial Mantel matrix tests and (ii) multiple regression on phylogenetic and geographic nearest neighbours. Mantel tests showed that both geographic and phylogenetic transmission was correlated with cultural diversity. Geographic distance had a greater overall partial correlation with cultural distance than did phylogenetic distance, but only phylogenetic correlations were found with kinship/social traits. Multiple regression on individual traits found that phylogenetic nearest neighbours predicted more cultural traits, especially those involving the inheritance of resources. Ancestral states of kinship traits were reconstructed using a Bayesian comparative method on a sample of 1000 phylogenies. The root of the tree was reconstructed as having matrilocal post-marital residence and a bilateral, flexible descent system. Proto Oceanic was reconstructed as unilocal and unilineal, and an hypothesis of matriliny and matrilocality could not be rejected. Murdock's main-sequence theory of the co-evolution of post-marital residence and descent systems was tested. The most likely model of the evolutionary pathway demonstrated that residence changed before descent. Rates of change in residence and descent traits were estimated. A co-evolutionary hypothesis of matriliny and male absence was tested. Contrary to anthropological theory, a high dependence on fishing showed no clear pattern of co-evolution with matrilineal social organisation. Population size of the language community was hypothesized to be a factor influencing lexical change. Conventional statistics showed a significant strong inverse correlation, indicating a relationship between small populations and accelerated lexical change. This correlation disappeared when comparative methods were used to control for phylogeny. Population size appeared to be evolving according to a drift model, while lexical change did not fit a neutral model of evolution.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:719058 |
| Date | January 2007 |
| Creators | Jordan, F. M. |
| Publisher | University College London (University of London) |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://discovery.ucl.ac.uk/1446081/ |
Page generated in 0.0026 seconds